Table 6.
IV Impacts of the Dissimilarity Index
| ln(Public Enrollment) |
ln(Private Enrollment) |
ln(Pop- ulation) |
|
|---|---|---|---|
|
|
|||
| 1 | 2 | 3 | |
| Panel A: Whites | |||
| Dissimilarity Index | 0.73** (0.33) |
0.19 (0.50) |
0.49* (0.28) |
| Dissimilarity Index * South | 0.99* (0.52) |
0.86 (0.67) |
0.96** (0.43) |
| Dissimilarity Index * Non-South | 0.30 (0.25) |
−0.97*** (0.33) |
−0.33 (0.26) |
| Panel B: Blacks | |||
| Dissimilarity Index (5+) | −0.40** (0.17) |
1.63*** (0.46) |
−0.14 (0.13) |
| Dissimilarity Index (5+) * South | −0.03 (0.14) |
2.02*** (0.75) |
0.01 (0.19) |
| Dissimilarity Index (5+) * Non-South | −0.76*** (0.24) |
1.31*** (0.48) |
−0.28 (0.19) |
| MSA & Year-South & West FE | X | X | X |
Note: Entries give coefficients and standard errors from IV regressions of central district outcomes listed in the column headings on the dissimilarity index. Within each column-panel combination, the results of two regressions are displayed: one which does not allow for regional heterogeneity and one that does permit regional heterogeneity. The desegregation indicator and the desegregation indicator interacted with an indicator for the West Census Region enter as instruments for the dissimilarity index in the first-stage (not shown). In Panel B, the dissimilarity index is measured 4 years prior to outcomes and is instrumented with the desegregation (5+) indicator used in Tables 4 and 5 and the desegregation (5+) indicator interacted with an indicator for the West Census region. For the regressions allowing regional heterogeneity, the segregation indicator is interacted with South and Non-South indicators. (The West interaction is also included and there are therefore three instruments and two first-stages in these specifications.) Standard errors are clustered at the MSA level. There are 313 observations in Panel A and 305 observations in Panel B. See the Data Appendix for information on the construction of the dissimilarity index variable. The first-stage F-statistics are 10.0 (Panel A, no regional heterogeneity), 9.1 (Panel A, South), 11.4 (Panel A, Non-South), 16.8 (Panel B, no regional heterogeneity), 10.8 (Panel B, South) and 11.4 (Panel B, Non-South).